Auto-organization and Emergence of Shared Language Structure

Abstract

The principal goal of attempts to construct computational models of the emergence of language is to shed light on the kinds of processes that may have led to the development of such phenomena as shared lexicons and grammars in the history of the human species. Researchers who attempt to model the emergence of lexicons make a set of shared assumptions about the nature of the problem to be solved. First, there are constraints on what counts as a shared lexicon. A lexicon is a systematic set of associations (a mapping) between forms and meanings. Forms are patterns. Tokens of a form are physical structures that bear the pattern of a particular form. For example, words are forms in this sense. Each instance of a particular word is a token of that word because it bears the pattern (sequence of sounds or letters) of that word. Forms must be discriminable from one another. Meanings are generally taken to be mental structures which, on the one hand, shape agents’ interactions with a world of objects and, on the other hand, also shape agents’ interactions with forms.